Multisensory perceptual learning in virtual reality may facilitate transfer to untrained locations but is impacted by training procedures.

Visual training improves performance in visually-intact and visually-impaired participants, making it useful as a rehabilitation tool. An interesting question in rehabilitation is whether invocation of multisensory integration could increase training efficacy. To investigate, participants completed a 10-day training experiment wherein they repeatedly performed a 4-way global motion direction discrimination task. On each trial, participants were presented with a 5{degrees} diameter visual global motion stimulus placed in a gaze-contingent manner at 10{degrees} azimuth/elevation. The visual-only group was presented the unimodal visual stimulus. However, for the auditory/visual (AV) group, the visual stimulus was paired with a pulsed white-noise auditory cue moving along in a direction consistent with the horizontal component of the visual motion stimulus. Direction range thresholds (DRT) were computed daily. The motion direction discrimination learning transferred fully in both groups, regardless of the presence of a feature-based attention cue. However, the figure-ground segregation learning only fully transferred to untrained locations with the addition of the feature-based attention cue. These results speak to the multiple levels of processing on which perceptual learning can operate. PACS0000, 1111 2000 MSC0000, 1111